delivered the opinion of the Court

We granted the State’s Petition for Discretionary Review to determine whether the Confrontation Clause of the United States Constitution was violated by the admission of a drug analysis when only the reviewing analyst (not the testing analyst) testified. We hold that there was such an error and affirm the judgment of the Fifth Court of Appeals to remand this case for a new trial.

I

The appellant was arrested by a Dallas police officer who saw him and a companion with drugs and paraphernalia. He was indicted for possession with intent to deliver a controlled substance, cocaine.1

At trial, the State offered into evidence a one-page lab report. The relevant findings stated only, “The contents of four green ziplock bags was used for analysis. The hard, white material contained cocaine. The amount of cocaine found was 1.38 grams (62%). The total weight of the material, including adulterants or dilutants was 2.2 grams.” The report was signed by Jennifer Pinckard, the analyst, and Monica Lopez, the reviewer.

The State called Lopez, but not Pinck-ard, to testify. Lopez testified that she was a supervisor for the Southwestern Institution of Forensic Sciences (SWIFS), a laboratory that is independent of the Dallas County D.A.’s Office and the police department. Lopez explained that Pinck-ard had performed all the tests in this particular case, but Pinckard no longer worked for SWIFS. No evidence was offered as to why Pinckard had left the laboratory. Lopez said that, as the reviewer, she was to ensure that the lab’s policies and procedures were followed. Although she agreed with the State that she “basically double-checked everything that *636was done,” she did not clarify what that meant. There was no indication that she actually saw the tests being performed or participated in them. The appellant objected, alleging a violation of his Sixth Amendment right to confront witnesses against him. The trial court overruled his objection and admitted the report, the underlying physical evidence, and Lopez’s testimony that the substance was cocaine.

The Fifth Court of Appeals held that the trial court erred by admitting the drug analysis and Lopez’s testimony that the evidence was cocaine. After finding that the error was not harmless,2 the Court of Appeals reversed and remanded for a new trial because the State had no other admissible evidence of the substance’s identity and weight.3 The State filed a Petition for Discretionary Review, which we granted.

II

Under the Confrontation Clause of the Sixth Amendment of the United States Constitution, made applicable to the states through the Fourteenth Amendment,4 “in all criminal prosecutions, the accused shall enjoy the right ... to be confronted with the witnesses against him.” In Crawford v. Washington, the Supreme Court interpreted this to mean that “testimonial” evidence is inadmissible at trial unless the witness who made the testimonial statement either:

(1) takes the stand to be cross-examined or

(2) is unavailable and the defendant had a prior opportunity to cross-examine him.5

The prior opportunity to cross-examine in person is both a necessary and a dispositive requirement for the admission of testimonial statements under the Confrontation Clause.6 The Court warned that “under no circumstances” shall the defendant be deprived of “seeing the witness face to face, and ... subjecting him to the ordeal of cross-examination.”7 The very real difficulties and costs involved in making witnesses available at trial cannot trump this categorical requirement. The Constitution does not list exceptions, and so (reasoned the Court) the judiciary should not create them.8

While the exact contours of what is testimonial continue to be defined by the courts, such statements are formal and similar to trial testimony. In other words, testimonial statements are those “that were made under circumstances which would lead an objective witness reasonably to believe that the statement would be available for use at a later trial.”9

These principles have been applied to forensic reports. The Supreme Court has explicitly held that an analyst’s certification prepared in connection with a criminal *637investigation or prosecution (specifically, a report identifying a substance as cocaine) is testimonial and cannot be admitted without satisfying the requirements of the Confrontation Clause.10 These reports are formal and created for the sole purpose of establishing or proving a highly relevant fact to a criminal prosecution.11 However, the issue of exactly who is required to testify in connection with such a report has not been fully resolved.

Bullcoming v. New Mexico was a prosecution for aggravated DWI in which the State did not call the analyst who tested the defendant’s blood sample for alcohol. Instead, because the original analyst was on unpaid leave, the State called another analyst familiar with the laboratory’s testing procedures.12 The Supreme Court held that, because the report was the testimonial statement of the analyst who performed the tests, it could not be offered into evidence through the testimony of a different, “surrogate” witness.13

In reaching this decision, the Supreme Court explicitly rejected the lower court’s reasoning that the analyst was only interpreting machine-generated data and that the testimonial statements were therefore those of the machine. Rather, the original analyst needed to be cross-examined so that the defendant could explore “the particular test and testing process he employed ... [and] any lapses or lies on the certifying analyst’s part.”14 Even if the results in question involved no interprétation or discretion, the testifying reviewer could not verify that the results were properly generated. Further, the court explained that the defendant had a right to question why the testing analyst was on unpaid leave.15

Ill

This case is controlled by Bull-coming. There is no indication that the appellant had a pre-trial opportunity to cross-examine the analyst who tested the cocaine. Although Lopez, the testifying ■witness, was a supervisor who “reviewed” the original process, we cannot say, on this record, that she had personal knowledge that the tests were done correctly or that the tester did not fabricate the results. She could say only that the original analyst wrote a report claiming to have conformed with the required safeguards. Consequently, cross-examining her did not satisfy the appellant’s constitutional rights.16 While we cannot say that anything would have come from a cross-examination of the original analyst, the law does not “tolerate dispensing with confrontation simply because the court believes that questioning one witness about another’s testimonial statements provides a fair enough opportunity for cross-examination.”17

The State argues that this case is distinguishable from Bullcoming because Lopez actually signed the report that was admitted into evidence. This is irrelevant. Without having the testimony of the ana*638lyst who actually performed the tests, or at least one who observed their execution, the defendant has no way to explore the types of corruption and missteps the Confrontation Clause was designed to protect against. It would not, for instance, solve the problem if a laboratory had all of its analysts sign every report. Rather, the witness being called needs to have personal knowledge of the facts in issue — the specific tests and their execution.18

The State also emphasizes parts of Justice Sotomayor’s Bullcoming concurrence in which she highlighted the limitations of the majority’s opinion. Specifically, Justice Sotomayor clarified that Bullcoming was “not a case in which the person testifying is a supervisor, reviewer, or someone else with a personal, albeit limited, connection to the scientific test at issue.”19 The State argues that this portion of Justice Sotomayor’s opinion indicates that a case such as this might have been decided differently by the Supreme Court. However, this is only one justice’s opinion, which does not have the weight of law even if it may indicate the Supreme Court’s changing course. This is especially true given that the State concedes that such an outcome is directly in conflict with the majority’s unequivocal statement that “the accused’s right is to be confronted with the analyst who made the certification, unless that analyst is unavailable at trial, and the accused had an opportunity, pretrial, to cross-examine that particular scientist.”20

IV

The Supreme Court handed down Williams v. Illinois21 the same day that the Fifth Court of Appeals handed down its decision in this case. Consequently, the Court of Appeals did not have the benefit of the most recent Supreme Court precedent. Williams was a splintered decision in which only an outcome, and not an opinion, received a majority vote. This made its full impact hard to discern. However, none of the plurality’s rationales affect the result in this case.

In Williams, semen samples taken from a rape victim’s vagina were sent to an independent laboratory for DNA testing. The lab created a DNA profile from the semen and sent that profile back to police. At trial, an expert for the prosecution testified that the profile created from samples in the victim’s vagina matched the defendant’s DNA profile generated by the police laboratory. The defendant argued that the expert could not testify that the DNA profile was created from semen within the victim because she only read a report generated by the independent lab.22 He further objected that his Sixth Amendment rights were violated when he was unable to cross-examine the analyst who created the profile, even though it was not offered into evidence.23

Justice Alito delivered the judgment of the court and an opinion that garnered four votes.24 In the view of the plurality, the petitioner’s rights were not violated because the out-of-court statement (that the DNA profile came from semen within the victim) was not offered to prove the truth of the matter asserted. Instead, it was merely alluded to in order to explain *639the basis of the witness’s independent conclusions (that the profile matched the petitioner’s DNA).25 Alternatively, as a second and independent basis for the decision, Justice Alito stated that the report was not testimonial because it was created before there was a specific suspect. Consequently, it was not inherently inculpatory or created for use against the petitioner.26

Neither of these rationales helps the State in this case. The statement at issue here was offered into evidence for its truth. The SWIFS report was submitted as the only evidence of the amount and makeup of the substance possessed by the appellant, the cocaine. Without such evidence, it is doubtful that the appellant could have been convicted at all. Justice Alito’s second rationale is equally inapplicable to this case, which is unlike Williams because the lab report was generated after the appellant was arrested and with the sole purpose of proving his guilt.

Justice Alito’s opinion stressed that Williams was not a departure from other cases such as Bullcoming because, in Williams, the actual report was not offered into evidence while in the other cases it was.27 Again, more in line with Bull-coming, the report at issue here was offered and admitted into evidence. Consequently, it was not merely mentioned as an underlying basis of the expert’s opinion: the report itself was primary evidence.

Although no other members of the Court joined Justice Thomas’s opinion, the State asserts that we should follow its reasoning because it is the “narrowest” ground for the judgment reached.28 Justice Thomas disagreed with Justice Alito and argued that the report was offered for the truth of the matter it asserted. Nonetheless, Justice Thomas reasoned that there was no Sixth Amendment violation because the report was not formal or solemn enough to be testimonial.29

While it is true that the report in this ease does not contain an oath, affirmation, or certificate as desired by Justice Thomas, we find this distinction irrelevant. The State cannot' sidestep the Sixth Amendment merely by choosing less formal language. The report asserted that the substance was cocaine, was signed by the analyst who performed the tests (presumably to certify the veracity of the report’s contents), and then was signed again by a reviewer. The only rational conclusion is that this is a formal statement created in preparation for trial and is testimonial in nature. We are not convinced that the absence of an oath or the word “certify” is enough of a distinction to change the analysis from that of Bullcoming. Even Justice Thomas, despite his emphasis on the ceremonial, stated that “the Confrontation Clause reaches bad-faith attempts to evade the formalized process.”30 That is precisely our case. The State can not avoid a straightforward application of Bullcoming by adding the signature of a reviewer with no personal knowledge and omitting more formalized language. We are not persuaded.

*640y

The lower court was correct that this case closely mirrors the facts of Bullcom-ing. The State attempted to submit testimonial evidence that the appellant possessed cocaine without giving the appellant the opportunity to cross-examine the analyst who tested the cocaine and made the affirmation of its contents. Although the State did call the reviewing analyst at trial, that witness did not have personal knowledge of the testimonial facts being submitted. Consequently, she was not an appropriate surrogate witness for cross-examination. We affirm the Court of Appeals’ decision to reverse and remand to the trial court for further proceedings.

KELLER, P.J., concurred.

filed a concurring opinion, in which KELLER, P.J., and HERVEY, J., joined.

For the two reasons described below, I join only the Court’s judgment.

First, the majority misinterprets the State’s argument by characterizing its claims that Bullcoming v. New Mexico1 is distinguishable from the present ease solely because the testifying witness, Monica Lopez — Jennifer Pinkard’s (the testing analyst) supervisor — signed the drug analysis report. This was only a small piece of the State’s argument. The main thrust of the State’s argument is that in Bullcoming the testifying witness “neither observed nor reviewed [the testing analyst’s] analysis,” but in this case, the witness did “review” the testing analyst’s analysis.2 In Bullcoming, a state laboratory analyst testified about the testing performed by another analyst in the same lab. Despite being familiar with the laboratory’s testing procedures, the witness neither participated in nor observed the test on Bullcom-ing’s blood sample, and the lab report was admitted as a business record.3 As a result, the Supreme Court concluded Bull-coming’s confrontation rights were violated.

The State highlights Lopez’s testimony to support its claim that Lopez’s connection with the drug testing was greater than the testifying witness in Bullcoming. After Pinkard performed the testing, Lopez reviewed the report to ensure that the lab’s policies and procedures were followed. She further stated that she “double checked everything that was done” and signed the report as the “reviewer.” The State asserts in its brief, “As a signer, [Lopez] certified the results and correctness of the content of the report.”4 While Lopez’s testimony may not be ideally descriptive about what exactly double-checking another analyst’s work entails, the fact that she performed some level of review illustrates the factual differences between the present case and Bullcoming. The State acknowledges that Lopez did not stand over Pinkard’s shoulder and observe her performing the test. But as the State asserts, Bullcoming does not necessarily require this. It appears to leave room within the contours of the Confrontation Clause for a “reviewer” to testify in the testing analyst’s stead.5 The majority gives the State’s argument short shrift.

*641Second, the majority’s response to the State’s argument that the admitted report’s lack of formality or solemnity rendered it non-testimonial is a gratuitous, ad hominem attack on the State. Rebuffing the State’s argument, the majority quotes Justice Thomas’s concurring opinion in Williams v. Illinois6 — the authority the State relied upon in advancing its argument that the lab report was non-testimonial — in which he stated “the Confrontation Clause reaches bad-faith attempts to evade the formalized process.”7 The majority continues: “That is precisely our case. The State can not avoid a straightforward application of Bullcoming by adding the signature of a reviewer with no personal knowledge and omitting more formalized language.”8 A plain reading of these passages suggests that the Court believes that the State has somehow influenced the lab’s procedures in a bad-faith attempt to skirt the Confrontation Clause’s requirements. No basis is given for this serious accusation. This is wholly unsupported, and as such, has no place in an opinion from this Court. I certainly will not affix my name to an opinion containing such a statement.

filed a concurring opinion in which JOHNSON, J., joined.

I join only the Court’s judgment and write separately to express concern about what I perceive to be unfair criticism of the forensic-science community and a limitation on the use of forensic testimony that is far too narrow.

The majority suggests that Appellant’s right to confront the witnesses against him was violated when the original analyst was not called to the stand at Appellant’s insistence because there is an absence of evidence showing that the “tester did not fabricate the results” and the analyst’s report claimed only to have conformed with the required safeguards. Maj. Op. at 637-38. The opinion further states:

Without having the testimony of the analyst who actually performed the tests, or at least one who observed their execution, the defendant has no way to explore the types of corruption and missteps the Confrontation Clause was designed to protect against.

Id. Although I believe that Appellant’s confrontation right was violated in this case, the majority unfairly suggests that an unavailable analyst is an untrustworthy analyst. I disagree. It is important for a criminal defendant to be able to cross-examine an analyst who sponsors a report that inculpates the defendant and to have the opportunity to question an analyst about why he or she is no longer employed at the laboratory, but there may be other important evidentiary considerations to take into account. And the suggestion that scientists in the forensic-science community are untrustworthy merely because they are unavailable to testify, when there is no evidence to support that position, undermines confidence in a profession that is governed by rigorous protocols and certification procedures.

Also, I would suggest a broader reading of Bullcoming — one that might allow, for example, a second analyst, with an independent analysis, opinion, or judgment, to testify to results of laboratory testing depending on the circumstances of the analysis and the definition of “analyst” or “reviewer.” If the State can produce “another” who may have developed his or her own separate conclusion based on data supplied through testing (i.e., particular “testing” is really performed through machinery and analysts develop opinions *642from that data), I see no reason why that witness should be denied an opportunity to testify. If you have an independent analysis based on specific data that would be the same for each analyst, how much personal observation is required by the second analyst before he can testify?

With these questions, I concur in the Court’s judgment.